Fig. 1: Screening assays to identify MM risk variants for transcriptional activity.
From: Functional dissection of inherited non-coding variation influencing multiple myeloma risk

a Manhattan plot of the largest genome-wide association study on MM to date, a meta-analysis totaling 9974 MM cases and 247,556 controls of European ancestry11. The 23 indicated loci associate with MM at P < 5 × 10−8. The 11q13.3 (CCND1) locus specifically associates with risk of t(11;14)[IGH/CCND1] translocation MM56. Lead variants at each locus are detailed in Supplementary Table 1. b We employed MPRA to screen variants in high LD (r2 > 0.8) with MM lead variants for transcriptional activity. For each variant, we designed twelve 120-bp oligonucleotide sequences representing the reference and alternative alleles in six genomic contexts (positive and negative DNA strand × three sliding windows with the variant positioned at −20, 0, or +20 bp from the center). The synthesized oligonucleotides were coupled to a synthetic reporter gene with 20-nt random sequence barcodes 3′ of its open reading frame. c Sequencing of the final plasmid library identified 1.73 × 106 barcodes mapping to 12,378 of the 12,468 designed oligonucleotides. The histogram shows the numbers of barcodes representing each oligonucleotide (median 103). d Following transfection of the library into cell lines, the transcriptional activity of each construct was measured by quantifying the barcode representation in reporter mRNA relative to DNA by sequencing. Barcode activity was quantified as log2(1+#RNAi)/(1+#DNAi) where #RNAi and #DNAi are the read counts for barcode i normalized to counts per 10 million reads. We performed MPRA in three replicates in each cell line. The plot shows the correlation of barcode activity estimates between two L363 replicates.